University Physics / Undergraduate Physics / Unit 15
Learning outcomes and scope
By the end of this unit, students should explain semiconductor carrier transport and junction behaviour; analyse diode, BJT and MOSFET circuits; design and evaluate op-amp amplifiers, feedback networks, oscillators and filters; interpret digital logic and timing; select sensors and signal-conditioning circuits; understand analogue-to-digital conversion, sampling, noise and grounding; and plan safe, traceable laboratory measurements with uncertainty budgets.
Constants and conventions: q=1.602176634×10⁻¹⁹ C; k_B=1.380649×10⁻²³ J K⁻¹; ε₀=8.8541878128×10⁻¹² F m⁻¹; V_T=k_BT/q≈25.9 mV at 300 K. Unless otherwise stated, equations are ideal first-order models; real-device limits must be checked against manufacturer data sheets.
Contents
- Semiconductor foundations and energy bands
- Intrinsic and extrinsic semiconductors
- Drift, diffusion and the Einstein relation
- The p–n junction and depletion region
- Diodes: rectification, clipping and protection
- Bipolar junction transistors: structure and action
- Transistor biasing and small-signal amplifiers
- Field-effect transistors and MOSFETs
- Operational amplifiers: ideal model and limitations
- Feedback, stability and oscillators
- Filters and frequency response
- Digital electronics and logic families
- Sensors and transducers
- Signal conditioning and instrumentation amplifiers
- Analogue-to-digital and digital-to-analogue conversion
- Noise, grounding and electromagnetic compatibility
- Power electronics and switching devices
- Oscilloscopes, multimeters and function generators
- Laboratory design: transistor and op-amp characterisation
- Reliability, thermal management and device protection
- Integrated measurement system and uncertainty budget
- Worked numerical examples
- Laboratory investigations
- Examination questions
1. Semiconductor foundations and energy bands
A crystalline semiconductor contains a periodic array of atoms whose outer electrons form covalent bonds. In a large crystal, the allowed electronic states form energy bands separated by forbidden gaps. The valence band is normally filled at low temperature and the conduction band is sparsely populated or empty. The band gap E_g separates the top of the valence band from the bottom of the conduction band. Silicon has an indirect gap, while gallium arsenide has a direct gap; this distinction strongly affects optical emission and absorption.
At finite temperature, some electrons gain enough thermal energy to cross the gap, leaving mobile holes in the valence band. An electron and a hole both contribute to electrical conduction, with their own mobilities and effective masses. The intrinsic carrier concentration depends strongly on temperature and band gap. Doping introduces donor or acceptor states near a band edge and changes the equilibrium carrier density by many orders of magnitude.
The conductivity is σ=q(nμ_n+pμ_p), where n and p are electron and hole concentrations and μ_n and μ_p are mobilities. Resistivity is ρ=1/σ. A material can therefore become more conductive either by increasing carrier concentration or by increasing mobility. In practical devices, doping concentration, temperature, defects, surfaces and geometry all influence the measured resistance.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
2. Intrinsic and extrinsic semiconductors
An intrinsic semiconductor is chemically pure enough that carriers are generated mainly as electron–hole pairs. Thermal equilibrium requires n=p=n_i. The mass-action law np=n_i² applies to a non-degenerate semiconductor at thermal equilibrium. The intrinsic concentration rises rapidly with temperature because the number of thermally excited carriers increases approximately with exp(−E_g/2k_BT), modified by effective density-of-states factors.
Donor impurities such as phosphorus in silicon provide an extra electron and create n-type material. Acceptor impurities such as boron create holes and p-type material. In n-type material electrons are majority carriers and holes minority carriers; in p-type material the reverse is true. Doping does not make a crystal electrically charged overall: ionised dopants are balanced by mobile carriers and the crystal remains approximately neutral in bulk.
At sufficiently high temperature, intrinsic excitation can dominate over dopant ionisation, and the material approaches intrinsic behaviour. At very high doping, impurity bands, band-gap narrowing and degeneracy may invalidate simple non-degenerate formulas. Always state whether the semiconductor is intrinsic, extrinsic, compensated or degenerate before applying carrier statistics.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
3. Drift, diffusion and the Einstein relation
An electric field produces drift motion. In a low-field regime, electron drift velocity is v_d=−μ_nE and hole drift velocity is v_d=μ_pE, with sign determined by charge. Conventional current points in the direction of positive charge flow, so electron motion is opposite to electron current. At high electric fields, carrier velocity may saturate and mobility is no longer a constant.
A concentration gradient produces diffusion from high concentration toward low concentration. Electron and hole diffusion current densities are proportional to their diffusion coefficients and concentration gradients. The drift–diffusion model combines field-driven and gradient-driven transport and is foundational for analysing junctions, transistors, photodiodes and solar cells.
At thermal equilibrium, drift and diffusion balance. For a non-degenerate semiconductor, the Einstein relation is D/μ=k_BT/q. It links diffusion coefficient D to mobility μ and thermal energy. The relation assumes near-equilibrium carrier statistics and is modified in degenerate materials. In device modelling, consistent sign conventions for electron and hole currents are essential.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
4. The p–n junction and depletion region
A p–n junction is formed by bringing p-type and n-type regions into intimate contact. Majority carriers diffuse across the interface: electrons move from n to p and holes from p to n. Recombination near the junction leaves behind fixed ionised donors and acceptors, creating a depletion region with an internal electric field. This field opposes further majority-carrier diffusion and establishes a built-in potential.
At equilibrium, the Fermi level is constant across the junction. The depletion approximation treats the mobile carrier density as negligible within the depletion region and the ionised dopants as fixed charge. Applying Poisson's equation dE/dx=ρ/ε and integrating gives a space-charge field and depletion width. The total depletion width depends on doping and applied bias; it extends farther into the more lightly doped side.
Forward bias reduces the junction barrier and allows a large diffusion current; reverse bias increases the barrier and leaves a small minority-carrier leakage current until breakdown. The ideal diode equation is I=I_S[exp(qV/nk_BT)−1]. Real devices show series resistance, recombination, leakage, temperature dependence and high-injection effects. The ideality factor n is often between 1 and 2 in practical regimes.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
5. Diodes: rectification, clipping and protection
A p–n diode conducts strongly in forward bias after its junction current becomes appreciable, while reverse current is comparatively small before breakdown. The often-quoted forward voltage of about 0.7 V for silicon is a practical approximation, not a universal fixed threshold; actual voltage depends on current, temperature and device construction. A diode's current–voltage curve is exponential over its ideal operating region.
Half-wave rectifiers use one diode to pass one polarity of an AC waveform. Full-wave rectifiers use a centre-tapped arrangement or a four-diode bridge to produce output pulses on both half-cycles. A smoothing capacitor charges near waveform peaks and discharges into the load between peaks. Ripple decreases when capacitance or supply frequency increases and load current decreases.
Diodes also serve as clamping and clipping elements, reverse-polarity protection, flyback paths for inductive loads and signal detectors. A Zener diode is designed to operate in reverse breakdown over a controlled region and can provide a reference voltage when current is limited appropriately. Check maximum reverse voltage, power dissipation, surge current and thermal conditions before selecting a device.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
6. Bipolar junction transistors: structure and action
A bipolar junction transistor has emitter, base and collector regions arranged as npn or pnp. The emitter is heavily doped to inject carriers, the base is thin and lightly doped to reduce recombination, and the collector is designed to collect carriers and dissipate heat. In forward-active operation, the emitter–base junction is forward biased and the collector–base junction reverse biased.
For an npn transistor, electrons injected from the emitter cross the thin base; most are swept into the collector. A small fraction recombines in the base, producing base current. The common-emitter current gain is β=I_C/I_B and the common-base gain is α=I_C/I_E, with α=β/(β+1). These are operating-point-dependent parameters rather than immutable constants.
A transistor can operate in cutoff, forward-active, saturation and reverse-active regions. In an amplifier, it is biased in the active region around a quiescent operating point. In a digital switch, it is driven between cutoff and saturation or used in a more efficient switching configuration. Base current, collector current, power dissipation, safe operating area and thermal runaway must be considered.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
7. Transistor biasing and small-signal amplifiers
Biasing establishes a stable DC operating point before an AC signal is applied. A voltage-divider network is commonly used to set base voltage, while an emitter resistor provides negative feedback that reduces sensitivity to transistor β and temperature. The Q-point should allow the intended signal swing without clipping into cutoff or saturation. Coupling capacitors pass AC while blocking DC; bypass capacitors can increase AC gain by reducing emitter degeneration over a selected frequency range.
For a small signal around the operating point, the transistor can be linearised. In the hybrid-π model, the transconductance is g_m=I_C/V_T, where V_T=k_BT/q is approximately 25.9 mV at room temperature. The input resistance r_π≈β/g_m. A common-emitter stage gives voltage gain with phase inversion; common-collector and common-base configurations offer different input/output impedances and gain properties.
Small-signal models are local approximations. They are valid when the signal excursion is sufficiently small that device characteristics can be linearised. Frequency response is limited by coupling and bypass capacitors at low frequencies and junction capacitances and transit time at high frequencies. Measure gain and phase over frequency rather than assuming a single mid-band gain applies everywhere.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
8. Field-effect transistors and MOSFETs
A field-effect transistor controls channel conductivity using an electric field. In a MOSFET, the gate is insulated from the semiconductor channel by a thin dielectric, giving very high DC input resistance. An n-channel enhancement MOSFET forms a conducting inversion channel when gate-to-source voltage exceeds threshold voltage V_TH. A p-channel device operates with opposite polarities.
In the long-channel square-law approximation, the saturation drain current is I_D≈(1/2)μC_ox(W/L)(V_GS−V_TH)², neglecting channel-length modulation and other non-idealities. In the linear region, the channel behaves approximately as a voltage-controlled resistor for small drain voltage. Modern short-channel devices depart from the simple square-law model due to velocity saturation, mobility degradation, threshold roll-off and other effects.
MOSFETs are widely used in digital logic, power conversion, analogue switching and sensor interfaces. Gate oxide can be damaged by electrostatic discharge, and maximum gate voltage must never be exceeded. Power MOSFET selection requires attention to on-resistance, gate charge, switching losses, safe operating area, body-diode behaviour and thermal resistance.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
9. Operational amplifiers: ideal model and limitations
An operational amplifier is a high-gain differential voltage amplifier with non-inverting and inverting inputs. The ideal model assumes infinite open-loop gain, infinite input resistance, zero output resistance, infinite bandwidth and zero input offset. With negative feedback and operation in the linear region, the input differential voltage is approximately zero and input currents are approximately zero. These are consequences of feedback, not physical short-circuit connections between the inputs.
The inverting amplifier has gain A_v=−R_f/R_in, while the non-inverting amplifier has gain 1+R_f/R_g. A voltage follower buffers a high-impedance source. A summing amplifier forms a weighted sum of input signals, and an integrator or differentiator uses capacitors in the feedback or input path. Practical circuits require frequency compensation and may need bias-current, offset and noise management.
Real op-amps have finite gain-bandwidth product, slew rate, input common-mode range, output swing, input offset voltage, bias current, noise and finite output current. The output saturates near the supply rails rather than reaching them exactly unless a rail-to-rail design is used. Check data-sheet limits and stability before connecting capacitive loads or using high closed-loop gain.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
10. Feedback, stability and oscillators
Negative feedback samples an output and subtracts a fraction from the input. For open-loop gain A and feedback factor β_f, closed-loop gain is A_cl=A/(1+Aβ_f). When loop gain is large, closed-loop gain depends mainly on the feedback network and becomes less sensitive to variations in A. Feedback can improve linearity and bandwidth, reduce distortion and stabilise gain, but it may introduce instability if phase shift and loop gain satisfy the oscillation condition.
The Barkhausen criterion for sustained sinusoidal oscillation is loop gain magnitude unity and total phase shift an integer multiple of 2π at the oscillation frequency. It is a necessary idealised condition, not a complete guarantee of stable startup or amplitude control. RC phase-shift, Wien-bridge, LC and crystal oscillators use different frequency-selective networks. Real oscillators require nonlinear amplitude limiting or automatic gain control.
Stability is assessed using loop gain, phase margin and gain margin. A circuit with insufficient phase margin may ring or oscillate. Decoupling, grounding, layout, load capacitance and component tolerances affect practical behaviour. In laboratory measurements, distinguish the oscillation frequency from harmonics and spurious modes, and state amplitude, distortion and frequency stability.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
11. Filters and frequency response
A filter selectively passes or attenuates frequency components. A first-order RC low-pass filter has transfer function H(jω)=1/(1+jωRC), with cutoff frequency f_c=1/(2πRC). A high-pass filter has complementary magnitude response. At cutoff, the magnitude is 1/√2 of the passband value, corresponding to −3.01 dB. The phase changes with frequency and should be included when timing or waveform shape matters.
Second-order filters provide steeper roll-off and can have Butterworth, Chebyshev, Bessel or other response characteristics. Butterworth maximises passband flatness; Chebyshev permits ripple for a sharper transition; Bessel prioritises group-delay behaviour and waveform preservation. Active filters use op-amps to provide gain and avoid inductors at low frequencies, but op-amp bandwidth and slew rate constrain performance.
Frequency response is measured by applying a sinusoidal input over a range of frequencies and recording output amplitude and phase. A Bode plot uses logarithmic frequency and decibel gain. For a voltage ratio, gain in decibels is 20log₁₀|V_out/V_in|. Keep source impedance, load impedance and instrument bandwidth in mind because they can alter the measured transfer function.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
12. Digital electronics and logic families
Digital circuits represent information using discrete logic levels. Binary variables take values conventionally labelled 0 and 1, but physical voltages occupy finite ranges and are subject to noise margins. NOT, AND, OR, NAND, NOR, XOR and XNOR gates form the basic combinational logic operations. De Morgan's laws allow transformations between AND/OR forms and are useful for implementing circuits using universal NAND or NOR gates.
Combinational circuits have outputs determined by present inputs. Examples include adders, multiplexers, decoders, encoders and comparators. Sequential circuits contain memory and depend on present inputs and previous states. Latches and flip-flops store bits; registers and counters are built from these elements. Clock timing, setup time, hold time, propagation delay and metastability determine reliable operation.
Logic families differ in supply voltage, noise margin, speed, power consumption and input/output current capability. CMOS is widely used because static power can be low, although dynamic switching power is approximately proportional to C_LV²f. Unused inputs should be tied to defined logic levels, and outputs must not be connected in ways that cause contention or exceed current ratings.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
13. Sensors and transducers
A sensor responds to a physical measurand such as temperature, pressure, displacement, light, magnetic field or chemical concentration. A transducer converts that response into a usable signal. Resistive temperature detectors use the temperature dependence of resistance; thermistors are strongly temperature-dependent resistors; thermocouples generate a voltage related to temperature difference; strain gauges change resistance with mechanical strain.
Sensor specifications include sensitivity, range, resolution, accuracy, precision, linearity, hysteresis, repeatability, drift, response time and cross-sensitivity. Sensitivity is the slope of output against measurand. A high sensitivity does not guarantee high accuracy: offset, calibration error, noise and environmental effects may dominate. A measurement chain should specify the sensor, excitation, signal conditioning, conversion and calibration method.
Choose a sensor by matching its range, response time, uncertainty and environmental rating to the experiment. Avoid loading a resistive sensor with a low input impedance. Use shielding and differential measurement for small signals, and consider grounding, cable capacitance, electromagnetic interference and temperature gradients. Calibration should span the intended operating range rather than relying on one reference point.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
14. Signal conditioning and instrumentation amplifiers
Many sensors produce low-level signals that require amplification, filtering, isolation or linearisation before digitisation. An instrumentation amplifier offers high input impedance, high common-mode rejection and accurately controlled gain. A differential amplifier rejects signals common to both inputs when resistor ratios are matched. Common-mode rejection ratio is CMRR=A_d/A_cm, often expressed as 20log₁₀(CMRR) dB.
Bridge circuits convert small resistance changes into differential voltages. A Wheatstone bridge is balanced when the ratio of one pair of arms equals the ratio of the other pair. Strain gauges and resistive sensors can be arranged in quarter-, half- or full-bridge configurations to improve sensitivity and temperature compensation. Excitation voltage must be limited to avoid self-heating.
Signal conditioning also includes anti-alias filters, level shifting, current-to-voltage conversion, isolation and protection. Select gain so the signal uses the converter's range without clipping. Check amplifier input common-mode limits and output swing, especially with single-supply operation. A carefully designed analogue front end often determines the quality of the entire measurement system.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
15. Analogue-to-digital and digital-to-analogue conversion
An analogue-to-digital converter samples a continuous signal in time and quantises its amplitude. For a converter with N bits over a full-scale range V_FS, the ideal quantisation step is approximately Δ=V_FS/2^N. Quantisation error is often modelled as uniformly distributed between ±Δ/2 when the signal sufficiently exercises the code range. The ideal full-scale sine-wave signal-to-quantisation-noise ratio is approximately 6.02N+1.76 dB.
The sampling theorem requires a sampling frequency greater than twice the highest frequency component for a band-limited signal, but practical systems need an anti-alias filter and margin above the Nyquist frequency. Aliasing folds out-of-band signals into the sampled band and cannot generally be removed by later digital processing. Clock jitter, aperture uncertainty, non-linearity, offset and gain error further limit converter performance.
A digital-to-analogue converter reconstructs an analogue output from digital codes using weighted currents, voltages or switched elements. Reconstruction filters suppress images created by the sampling process. Resolution, monotonicity, integral and differential non-linearity, settling time, glitch energy and reference stability are important specifications. Always distinguish nominal bit count from effective number of bits.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
16. Noise, grounding and electromagnetic compatibility
Noise may be thermal, shot, flicker, burst, quantisation or externally coupled interference. A resistor at temperature T has mean-square thermal noise voltage 4k_BTRΔf over bandwidth Δf. Shot noise arises from discrete charge transport and has mean-square current 2qIΔf under the usual assumptions. Flicker noise often dominates at low frequencies and depends on device technology and bias.
Signal-to-noise ratio compares signal power with noise power and is commonly expressed as 10log₁₀(P_s/P_n) dB. For voltage measurements into equal impedances, it can be expressed using 20log₁₀(V_s/V_n). Reducing bandwidth can reduce integrated white noise, but filtering must not remove information needed from the signal. Averaging reduces uncorrelated random noise but does not remove systematic offsets or correlated interference.
Grounding and shielding should control return-current paths and prevent unintended loops. Use a single-point or carefully engineered ground strategy where appropriate, twisted pairs for differential signals and shield termination suited to frequency and safety requirements. Keep high-current switching paths away from sensitive analogue inputs. Electromagnetic compatibility includes both limiting emissions and maintaining immunity to external disturbances.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
17. Power electronics and switching devices
Power electronics controls electrical energy using semiconductor switches. A diode conducts in one direction, while power MOSFETs and insulated-gate bipolar transistors can be actively switched. A buck converter steps down DC voltage, a boost converter steps it up, and a buck–boost topology can produce a magnitude above or below the input depending on configuration. Inductors and capacitors transfer and store energy between switching intervals.
For an ideal continuous-conduction buck converter, V_out≈D V_in, where D is the duty ratio. For an ideal boost converter, V_out≈V_in/(1−D). These relations neglect switch drops, diode losses, inductor resistance, switching transitions and control-loop effects. Inductor current ripple, capacitor voltage ripple, switching frequency, efficiency and thermal design must be considered together.
Switching devices dissipate conduction and switching losses. Gate-drive voltage, gate charge, switching speed, dead time, reverse recovery and parasitic inductance influence performance. Snubbers and freewheel paths control transients. High-voltage circuits store hazardous energy in capacitors even after power is removed; verify discharge and use appropriate isolation and safety procedures.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
18. Oscilloscopes, multimeters and function generators
A digital oscilloscope samples voltage versus time and displays the waveform. Key specifications include analogue bandwidth, sample rate, record length, vertical resolution, input impedance and probe attenuation. A probe can load a circuit and alter its behaviour; a compensated 10× probe usually offers lower capacitance than a 1× setting. The instrument's bandwidth limits the observed rise time and high-frequency amplitude.
A digital multimeter measures voltage, current, resistance and other quantities using internal conversion and range selection. Input impedance, burden voltage, fuse rating, measurement category and uncertainty vary by function. Never place a current input directly across a voltage source. A function generator supplies waveforms with specified amplitude, offset, frequency and output impedance; amplitude settings may assume a particular termination.
Instrument readings should be checked against calibration status and operating range. Record whether amplitude means peak, peak-to-peak or RMS. For a sinusoid, V_RMS=V_peak/√2. Use the correct coupling, trigger mode, probe factor and termination. Mains-referenced oscilloscope grounds can create short circuits if attached carelessly; use approved differential or isolated measurement methods where required.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
19. Laboratory design: transistor and op-amp characterisation
A transistor characterisation practical measures input and output characteristics at controlled bias. For a BJT, vary base current and collector-emitter voltage while recording collector current, then identify cutoff, active and saturation regions. Determine current gain at several operating points and observe how temperature changes the characteristic. Limit supply current and use suitable resistors to protect the device.
An op-amp practical can measure closed-loop gain, input and output resistance, bandwidth, slew rate and saturation limits. For a non-inverting amplifier, compare measured gain with 1+R_f/R_g. Sweep frequency to locate the −3 dB point and record phase response. Apply a large-amplitude square wave to estimate slew rate from the maximum output slope, while avoiding overload beyond data-sheet limits.
A laboratory report should include circuit diagrams with component values and supply rails, instrument settings, raw tables, uncertainty estimates, calculated values, plots and residual analysis. Explain discrepancies through component tolerances, loading, finite gain-bandwidth, parasitic capacitance, wiring and measurement uncertainty rather than simply labelling them experimental error.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
20. Reliability, thermal management and device protection
Electronic components have maximum voltage, current, power and temperature ratings. Junction temperature is estimated from ambient temperature plus dissipated power multiplied by the relevant thermal resistance. A device can fail even when average power appears acceptable if transient safe-operating-area limits are exceeded. Heat sinks, thermal interface materials, airflow and board copper reduce thermal resistance.
Protection includes current limiting, fuses, transient-voltage suppressors, flyback diodes, reverse-polarity protection, ESD controls and thermal shutdown. A fuse protects wiring and equipment against specified overcurrent conditions; it is not a substitute for correct circuit design. ESD-sensitive MOS gates and integrated circuits should be handled using grounded mats, wrist straps and suitable packaging.
Reliability engineering considers component derating, environmental stress, vibration, humidity, corrosion, thermal cycling and manufacturing variation. Design margins should follow component data sheets and applicable standards. A fault analysis asks how a single component failure affects the system and whether the resulting state is safe. Document test conditions and never exceed a device's absolute maximum ratings.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
21. Integrated measurement system and uncertainty budget
A complete measurement system comprises a measurand, sensor, excitation, analogue conditioning, filter, ADC, software, calibration and reporting. Each stage can contribute offset, gain error, nonlinearity, noise, drift and timing uncertainty. The system should be designed from the required measurement uncertainty backwards: determine the needed sensor range and sensitivity, allowable analogue noise, converter resolution and calibration interval.
For independent standard uncertainties u_i in a result y=f(x_i), first-order propagation gives u_y²=Σ(∂f/∂x_i)²u_i². Correlated inputs require covariance terms. Repeated observations estimate Type A uncertainty, while calibration certificates, resolution, manufacturer specifications and environmental limits contribute Type B estimates. A complete budget identifies each contribution, distribution, sensitivity coefficient and combined result.
Validation uses reference standards, repeat measurements, linearity checks, drift monitoring and independent comparison. Preserve raw data and analysis code, record instrument identifiers and calibration dates, and distinguish corrections from uncertainties. A result should be traceable through an unbroken calibration chain to recognised standards where possible.
Design and analysis practice. Begin by drawing the circuit or device cross-section and labelling all node voltages, current directions, polarities and reference grounds. State whether the model is DC, small-signal AC, transient or frequency-domain. Identify assumptions such as ideal diode, long-channel MOSFET, infinite op-amp gain or negligible loading. Solve symbolically before substituting values, then check units, limiting cases and power balance. Compare the calculated operating point with supply rails and component ratings before constructing a physical circuit.
Practical limitations. Real components have tolerances, parasitic resistance and capacitance, finite bandwidth, temperature dependence and noise. A measured response can differ from a simple prediction because the probe, generator, breadboard, cable or load changes the circuit. Distinguish a model's approximation error from measurement uncertainty and component variation. Use manufacturer data sheets for pinouts, absolute maximum ratings, thermal resistance and safe operating area; never infer these limits from a generic textbook diagram.
Worked numerical examples
Each result uses an explicitly stated idealised model. In practical design, include tolerances, temperature, loading, bandwidth and device ratings.
Example 1. Intrinsic conductivity
Given n=p=1.0×10¹⁶ m⁻³, μ_n=0.14 m²V⁻¹s⁻¹ and μ_p=0.05 m²V⁻¹s⁻¹, σ=q(nμ_n+pμ_p)=1.602×10⁻¹⁹×1.9×10¹⁵≈3.04×10⁻⁴ S m⁻¹.
Example 2. Einstein relation
At 300 K, D/μ=k_BT/q≈0.02585 V. For μ=0.12 m²V⁻¹s⁻¹, D≈3.10×10⁻³ m²s⁻¹.
Example 3. Diode current
For I_S=2 nA, V=0.60 V, n=1 and T=300 K, qV/kT≈23.2. I≈I_S exp(23.2)≈23.8 mA, illustrating the exponential sensitivity; real series resistance limits the ideal prediction.
Example 4. Bridge balance
A Wheatstone bridge balances when R₁/R₂=R₃/R₄. If R₁=120 Ω, R₂=100 Ω and R₃=240 Ω, then R₄=200 Ω.
Example 5. BJT current gain
If I_C=4.8 mA and I_B=40 μA, β=I_C/I_B=120. Emitter current is approximately 4.84 mA.
Example 6. MOSFET ideal saturation current
For μCox(W/L)=2 mA V⁻² and overdrive V_OV=1.5 V, I_D≈½×2×1.5²=2.25 mA, neglecting channel-length modulation.
Example 7. Inverting amplifier
With R_in=10 kΩ and R_f=100 kΩ, A_v=−10. An input sine amplitude of 0.20 V gives an ideal output amplitude of 2.0 V with 180° phase inversion, if not clipped.
Example 8. RC cutoff
For R=10 kΩ and C=10 nF, f_c=1/(2πRC)=1.59 kHz.
Example 9. ADC resolution
A 12-bit ADC over 0–3.3 V has Δ≈3.3/4096=0.806 mV per code. Ideal quantisation uncertainty for one reading is often modelled as Δ/√12≈0.233 mV standard uncertainty.
Example 10. Sampling and aliasing
A 1.2 kHz sinusoid sampled at 2.0 ksample s⁻¹ appears at 0.8 kHz in the first Nyquist band because it folds about 1.0 kHz. An anti-alias filter must attenuate components above the usable band.
Example 11. Thermal noise
For R=10 kΩ, T=300 K and Δf=10 kHz, v_n=√(4kTRΔf)≈1.29 μV RMS.
Example 12. Buck converter
For ideal V_in=24 V and duty ratio D=0.40, V_out≈9.6 V in continuous conduction.
Laboratory investigations and practical methods
1. Diode characteristics and rectification
Connect a silicon diode in series with a current-limiting resistor and a current-limited DC supply. Sweep forward voltage in safe increments and record diode voltage and current. Reverse the diode only with a suitably limited supply and remain well below its rated reverse voltage. Plot I against V using linear and semi-log axes. Construct a bridge rectifier with an isolated low-voltage AC source, observe the output with an oscilloscope and compare ripple with and without a smoothing capacitor. Record diode drops, load current, ripple frequency and uncertainty.
2. BJT common-emitter characteristics
Build a common-emitter test circuit using a low-voltage supply and resistors chosen to limit current. At several fixed base currents, vary collector-emitter voltage and record collector current. Plot output curves and identify cutoff, active and saturation regions. Calculate β at selected active-region points. Repeat at a different temperature only under controlled conditions and observe the shift in operating point. Do not exceed transistor dissipation or safe operating area.
3. Op-amp gain and frequency response
Build an inverting and a non-inverting amplifier using a suitable low-voltage op-amp supply. Measure closed-loop gain at a mid-band frequency, then sweep frequency and record magnitude and phase. Determine the −3 dB bandwidth and compare the gain-bandwidth product with the data sheet. Apply a square wave and observe slew-rate limitation and output saturation. Keep common-mode input and output swing within ratings, and use a decoupling capacitor close to the supply pins.
4. RC filter and Bode plot
Construct a first-order low-pass filter and measure input and output amplitudes over a logarithmic frequency sweep. Calculate gain in dB as 20log₁₀(V_out/V_in). Identify the cutoff where gain has fallen by approximately 3 dB from the passband. Compare measured cutoff with 1/(2πRC). Repeat with a high-pass configuration. Record generator output impedance and oscilloscope probe loading because both can alter the effective circuit constants.
5. Sensor calibration and bridge measurement
Connect a temperature-dependent resistor or strain gauge in a bridge. Use traceable reference points across the working range and record output voltage at each point. Fit an appropriate linear or nonlinear calibration curve and inspect residuals. Estimate sensitivity, hysteresis and repeatability. For small bridge outputs, use a differential or instrumentation amplifier and check common-mode range. Avoid sensor self-heating by limiting excitation power.
6. ADC sampling and aliasing demonstration
Apply a sinusoidal signal to an ADC input through an anti-alias filter. Sample at several rates above and below twice the signal frequency, then plot the reconstructed samples. Demonstrate apparent lower-frequency aliasing when the sampling rate is insufficient. Increase sampling rate and compare amplitude quantisation at different converter resolutions. Record clock source, sample rate, analogue bandwidth, full-scale range and any clipping.
7. Noise measurement
Terminate a low-noise amplifier input with a known resistance and measure output noise over a defined bandwidth. Compare integrated RMS noise with the thermal-noise prediction after accounting for amplifier voltage and current noise and gain. Change measurement bandwidth and observe the integrated noise change. Use shielding, short leads and suitable grounding; keep mains interference and switching supplies away from the sensitive input.
Laboratory safety and reporting
Use isolated low-voltage supplies for student experiments unless a specifically authorised high-voltage facility is provided. Check polarity and pinouts before power is applied. Discharge capacitors safely, observe ESD precautions, use current limits and never connect an earth-referenced oscilloscope ground to a live node without an approved differential measurement arrangement. Reports should include circuit schematics, component values, raw data, calibration status, instrument ranges, uncertainty budget, plots, residuals, conclusions and limitations.
Core equations and design checklist
Conductivity: σ=q(nμ_n+pμ_p)
Einstein relation: D/μ=k_BT/q
Diode: I=I_S[exp(qV/nk_BT)−1]
BJT gains: β=I_C/I_B; α=I_C/I_E
MOSFET saturation (long-channel ideal): I_D≈½μC_ox(W/L)(V_GS−V_TH)²
Inverting op-amp: A_v=−R_f/R_in
Non-inverting op-amp: A_v=1+R_f/R_g
RC cutoff: f_c=1/(2πRC)
ADC step: Δ≈V_FS/2^N
Ideal ADC sine SQNR: 6.02N+1.76 dB
Thermal noise: v_n=√(4k_BTRΔf)
Ideal buck: V_out≈DV_in; ideal boost: V_out≈V_in/(1−D)
Uncertainty: u_y²=Σ(∂f/∂x_i)²u_i² for independent inputs
Common errors to avoid
- Confusing electron flow with conventional current direction.
- Treating 0.7 V as an exact silicon diode threshold.
- Using ideal diode or transistor equations outside their operating regions.
- Ignoring transistor bias stability and thermal dissipation.
- Assuming an op-amp output can reach its supply rails or has unlimited bandwidth.
- Applying the Barkhausen criterion as a complete stability proof.
- Forgetting source and load impedance when measuring a filter.
- Leaving digital inputs floating or ignoring setup and hold time.
- Assuming ADC bit count equals effective resolution.
- Ignoring aliasing, grounding, probe loading and instrument calibration.
- Connecting an oscilloscope earth clip to an unsafe or live circuit node.
University examination questions with model answers
Show circuit diagrams, define reference directions, state assumptions and include units. Where a design is requested, justify component values and operating limits.
1. Explain intrinsic and extrinsic semiconductors.
Model answer: Intrinsic material has thermally generated electron–hole pairs with n=p. Extrinsic material is doped with donors or acceptors so one carrier type is the majority carrier.
2. Derive the conductivity expression.
Model answer: Electron and hole drift contributions add: J=q n μ_n E+q p μ_p E. Comparing J=σE gives σ=q(nμ_n+pμ_p).
3. State the Einstein relation and its assumptions.
Model answer: D/μ=k_BT/q for non-degenerate carriers in thermal equilibrium under the usual drift–diffusion assumptions.
4. Explain depletion-region formation.
Model answer: Majority carriers diffuse across the junction and recombine, leaving fixed ionised dopants. The resulting space-charge field opposes further diffusion.
5. Sketch and explain a diode I–V curve.
Model answer: Forward current rises exponentially; reverse current is small until breakdown. Mark forward region, leakage and breakdown, noting temperature and series-resistance effects.
6. Compare half-wave and bridge rectifiers.
Model answer: A half-wave circuit uses one diode and produces one pulse per cycle. A bridge uses four diodes and rectifies both half-cycles, with two diode drops in the conducting path.
7. Explain BJT active operation.
Model answer: The emitter-base junction is forward biased and collector-base junction reverse biased. Carriers injected from emitter cross the thin base and are collected.
8. Why is transistor bias stabilisation necessary?
Model answer: Device gain and junction characteristics vary with temperature and device spread. Negative feedback in the bias network stabilises the quiescent operating point.
9. Compare BJT and MOSFET control.
Model answer: A BJT is controlled by carrier injection and base drive; a MOSFET is controlled by gate electric field and has very low steady-state gate current.
10. Derive the inverting op-amp gain.
Model answer: With virtual ground, input current V_in/R_in flows through R_f. Thus V_out=−(R_f/R_in)V_in.
11. What is gain-bandwidth product?
Model answer: For a compensated op-amp, closed-loop gain and bandwidth trade approximately as a constant product over the applicable range.
12. State the Barkhausen criterion and its limitation.
Model answer: Loop gain magnitude is unity and total phase shift is 0° modulo 360° at oscillation. It is not alone sufficient to establish startup or stable amplitude.
13. Define the −3 dB cutoff frequency.
Model answer: It is the frequency where power falls to half its passband value and voltage magnitude to 1/√2 for equal impedance.
14. Explain setup and hold time.
Model answer: Data must be stable for specified intervals before and after the active clock edge for a flip-flop to capture reliably.
15. Define sensor sensitivity and resolution.
Model answer: Sensitivity is output change per unit measurand; resolution is the smallest change distinguishable by the measurement system.
16. What is CMRR?
Model answer: The ratio of differential gain to common-mode gain, usually expressed in decibels; high CMRR rejects interference common to both inputs.
17. Explain aliasing.
Model answer: Sampling maps frequencies separated by integer multiples of sampling frequency onto the same discrete-time frequency. Anti-alias filtering is required before conversion.
18. Calculate ideal ADC quantisation step.
Model answer: For N bits over full-scale range V_FS, Δ≈V_FS/2^N. Explain that effective resolution is reduced by noise and converter non-idealities.
19. Compare thermal and shot noise.
Model answer: Thermal noise arises from random thermal carrier motion and scales with resistance, temperature and bandwidth; shot noise arises from discrete charge flow and scales with average current and bandwidth.
20. Why is grounding important?
Model answer: Return-current paths and ground potential differences can inject interference, create loops or compromise safety. Grounding strategy must suit signal frequency and installation.
21. Explain MOSFET switching losses.
Model answer: Losses arise during voltage-current overlap in transitions, gate charging, conduction resistance, output capacitance and diode recovery.
22. What does oscilloscope bandwidth specify?
Model answer: The frequency at which the measured sinusoidal amplitude is attenuated by the specified amount, conventionally −3 dB.
23. How is uncertainty propagated through a measurement chain?
Model answer: Use partial derivatives and covariance terms for the measurement model, combine independent standard contributions and report the budget.
24. Distinguish accuracy and precision.
Model answer: Accuracy describes closeness to a reference value; precision describes agreement among repeated measurements.
25. List essential laboratory electrical safeguards.
Model answer: Use current-limited supplies, correctly rated probes and fuses, insulated connections, safe discharge of capacitors, appropriate isolation and never defeat protective earth or interlocks.
Advanced circuit analysis and design derivations
1. Deriving the diode small-signal resistance
For an ideal diode in forward conduction, I=I_S exp(V_D/nV_T) when the reverse term is negligible. Differentiate current with respect to voltage at a DC operating point I_D: dI/dV=I_D/(nV_T). The incremental resistance is therefore r_d=dV/dI=nV_T/I_D. At room temperature, a diode carrying 1 mA has an ideal incremental resistance near 25.9 Ω for n=1, or about twice that for n=2. This is not the same as the DC ratio V_D/I_D. The small-signal resistance is valid only for sufficiently small perturbations around the selected bias current. Series resistance and high-level injection increase the measured value at larger currents.
2. Deriving the depletion width of an abrupt junction
For an abrupt junction with acceptor concentration N_A on the p side and donor concentration N_D on the n side, charge neutrality requires N_Ax_p=N_Dx_n. Within the depletion approximation, Poisson's equation is dE/dx=ρ/ε_s. Integrating the piecewise constant charge density gives a linearly varying electric field in each depleted region. Integrating the field again gives the electrostatic potential. The total depletion width is W=x_p+x_n=[2ε_s(V_bi−V_A)/q(1/N_A+1/N_D)]¹ᐟ² for the stated bias convention, where forward bias V_A reduces the barrier. The lightly doped side contains the larger share of the depletion width. This relation is used to estimate junction capacitance and breakdown behaviour, but assumes uniform doping, one-dimensional geometry and a depletion region with negligible mobile charge.
3. Deriving the op-amp closed-loop gain
Let the open-loop relation be V_o=A(V_+−V_−), and let a fraction β_f of the output be returned to the inverting input. For a non-inverting amplifier, V_−=β_fV_o and V_+=V_i. Substitution gives V_o=A(V_i−β_fV_o), so V_o/V_i=A/(1+Aβ_f). When Aβ_f is much greater than one, the closed-loop gain approaches 1/β_f. The feedback network therefore determines gain and reduces sensitivity to finite open-loop gain. If the loop gain falls with frequency, the approximation eventually fails and phase shift can threaten stability. This derivation makes clear why negative feedback improves predictable gain only within the amplifier's stable bandwidth.
4. Deriving the first-order RC transfer function
For a resistor R in series with a capacitor C to ground and output measured across the capacitor, the impedances are Z_R=R and Z_C=1/(jωC). The voltage divider gives H(jω)=Z_C/(R+Z_C)=1/(1+jωRC). Its magnitude is |H|=[1+(ωRC)²]⁻¹ᐟ² and phase is −tan⁻¹(ωRC). At ω_c=1/RC, the magnitude is 1/√2 and the phase is −45°. Above cutoff, the asymptotic magnitude falls by approximately 20 dB per decade. The actual cutoff shifts if source resistance or load resistance is not negligible; include these in the equivalent network before selecting component values.
5. Quantisation uncertainty and converter performance
An ideal uniform quantiser rounds an input to the nearest code. If the quantisation error is uncorrelated with the signal and uniformly distributed from −Δ/2 to +Δ/2, its mean is zero and variance is Δ²/12. The standard uncertainty is therefore Δ/√12. This statistical model is useful for noise budgeting but can fail for slowly varying or periodic signals that do not sufficiently exercise converter codes. Dither can decorrelate quantisation error in some applications, at the cost of added noise. Real ADC performance also includes integral nonlinearity, differential nonlinearity, missing codes, aperture jitter, reference noise and front-end distortion.
6. Thermal design from junction temperature
For a device dissipating power P_D, a simple steady-state estimate is T_J=T_A+P_Dθ_JA, where θ_JA is junction-to-ambient thermal resistance for the stated mounting and airflow. If a heat sink is used, the thermal path is modelled as junction-to-case, case-to-sink interface and sink-to-ambient resistances in series. The allowable power is constrained by the maximum junction temperature and the actual ambient environment. Transient pulses require transient thermal impedance rather than the steady-state formula. Derating curves and safe operating area in the manufacturer's data sheet take precedence over a generic calculation.
7. Measurement bandwidth and integrated noise
For white noise with approximately constant spectral density over a bandwidth B, integrated mean-square noise is proportional to B and RMS noise is proportional to √B. A low-pass filter can therefore reduce broadband noise by limiting measurement bandwidth. However, the equivalent noise bandwidth of a real filter is not always equal to its nominal −3 dB bandwidth. For precision work, integrate the squared magnitude response over frequency or use the specified equivalent noise bandwidth. Report the bandwidth and weighting used whenever quoting RMS noise; two measurements with different bandwidths are not directly comparable.
8. Design review: from requirement to verified circuit
A disciplined design begins with measurable requirements: input range, output range, bandwidth, load, accuracy, power, environmental conditions and safety constraints. Translate each requirement into component and architecture choices. Perform hand calculations, simulate the circuit with realistic models, inspect worst-case tolerances and thermal limits, then build a prototype. Verify each functional block independently before system integration. Compare measured performance with acceptance limits and document deviations. A circuit that works once on a bench is not necessarily robust: repeat tests across supply, temperature, load and component variation, and consider failure modes and safe behaviour.
Integrated design case study: temperature data acquisition
Consider a system that measures temperature from a resistive sensor over a defined operating range and sends the result to a microcontroller. First select the sensor from its range, sensitivity, interchangeability, response time and long-term stability. A platinum resistance sensor offers a relatively predictable resistance–temperature relation, while a thermistor offers higher sensitivity over a narrower range. The excitation circuit must be chosen to provide adequate signal while limiting self-heating. A constant-current source can simplify conversion of resistance to voltage, but its compliance voltage and noise must be checked.
Next design the analogue front end. Estimate the smallest sensor voltage change corresponding to the required temperature resolution. Choose amplifier gain so the expected signal occupies a useful fraction of the ADC input range without clipping at either temperature extreme. Check input offset, bias current, common-mode range, reference stability and amplifier noise. Add an anti-alias filter whose passband preserves the sensor's real time variation while attenuating unwanted high-frequency interference. The ADC sampling rate should exceed twice the highest retained signal frequency, with practical margin and a defined anti-alias attenuation target.
Calibrate the complete chain rather than calibrating the sensor alone. Apply traceable temperature references at several points, record raw ADC codes and fit the conversion relation. Validate the model using independent check points and examine residuals for curvature or hysteresis. Build an uncertainty budget including sensor tolerance, calibration reference, repeatability, ADC quantisation, reference voltage, amplifier offset and temperature gradients. Store calibration coefficients with version information and retain raw readings for audit. Finally, test power interruption, open- and short-circuit sensor faults, out-of-range inputs and electromagnetic interference. A robust instrument reports invalid or uncertain conditions rather than silently presenting a plausible but incorrect temperature.
Further study
Unit 16: Experimental Physics II develops advanced electrical and optical measurement, calibration, uncertainty budgets, instrumentation and reproducible laboratory methods.
Continue to Unit 16: Experimental Physics II →